Should we assume that small enterprises in developing countries are lacking in business skills—and that guidance and training will improve these businesses? Economic theory says that firms do as much as possible to maximize profits—including paying for advice from management consultants. In developing countries, interventions ranging from quick lectures during microcredit meetings to extended engagements with international consulting firms aim to improve management practices. These interventions presume that the existing management must be missing something. And whenever there is a ton of activity, questionable data, and competing theories, researchers often try to fill the knowledge gap. We want to know: What is all this interventionist effort for? Can mere advice really help these enterprises run better, earn more money, and create more jobs—and, if so, why?
Two randomized evaluations recently conducted by Innovations for Poverty Action (IPA) in Ghana (by Dean Karlan, Ryan Knight, and Chris Udry) and in Mexico (by Miriam Bruhn, Dean Karlan, and Antoinette Schoar) explore this question for small and medium enterprises (SMEs). We aim to shed light on when advice will help SMEs, so that policymakers can decide how best to support them. The evaluations also challenge our assumptions.
In February, we asked SSIR website readers to predict the results of the two studies. Most got the answer wrong. Granted, the online article was short, with little room to provide the full contexts from which readers could make their predictions. Nearly everyone expected the consulting advice to affect enterprise growth positively in at least one of the studies, and almost half expected positive impact in both studies. The consulting advice did positively impact enterprises in Mexico, but generated some adverse effects in Ghana.
Why did we study these programs with randomized evaluations? For years, evaluations of such programs would follow participants over time, observing the change in their business afterward compared to beforehand. But countless outside factors influence the success, or failure, of a business. This fails to answer the question of impact: How have the businesses changed compared to how they would have changed had the training or consulting not taken place? Even if an analyst compares nonparticipants to participants, one must ask: Why are some participating and others not? Could it be that those participating are striving to improve their businesses? Random assignment to treatment and control addresses these factors.
In Accra, Ghana, IPA partnered with Ernst & Young to provide urban tailors and seamstresses with customized consulting advice on record keeping, customer service, and management of employees. Those offered the consulting services were randomly selected from a group of 160 tailors and seamstresses; we subsequently compared business outcomes for those selected to receive the consulting with those who got no consulting.
Our results showed that although the consulting intervention caused short-term changes in business practices, these impacts dissipated within a year after the consulting ended. On average, we found no long-term benefit from the consulting, and actually lower short-term profits. We believe some businesspeople hoped the advice would work and thus took it. But better bookkeeping and other business practices potentially took time away from the physical act of sewing clothes. Once profits took a hit, enterprise owners likely abandoned the practices and reverted to their previous methods.
If these tiny firms don’t benefit from consulting, would they benefit from more capital? To test this hypothesis, we handed out unconditional cash grants of $160—roughly equal to the businesses’ average monthly revenues—to a random subset of the tailors and seamstresses. The cash was generally invested in the businesses. As with the advice, the cash grants did not lead to increases in profits, but rather decreased profits. We see the capital infusion as not much different from the advice infusion: both represent a push from afar to expand operations, when these businesses were actually operating at their optimal scale. Following these interventions, the business owners made some changes, but they didn’t work out well. So they eventually reverted to former practices (which is good).
Note that 61 percent of respondents to our online survey assumed that the consulting in Ghana would raise firms’ profits. This corresponds with the common development practitioner and donor assumption that motivates advice-giving interventions for micro and small enterprises around the world. Our result does not mean, for the record, that no such program works. But it should make us think harder about the conditions necessary for success, and to test rigorously to see if we are right.
As for the Mexico study results, 77 percent of respondents to the SSIR online survey who thought we would find positive results from consulting in Mexico were right. We found an average 80 percent increase in sales and an average 120 percent increase in profits for firms receiving consulting advice, compared to the group that did not receive the intervention.
But the Mexico study was different from Ghana in that the program openly advertised to find interested firms. Those that responded were then entered into a lottery, and winners were matched with a local consultant who worked with the firm to define the scope of the consulting services. We must note, however, that we did not study the impact on firms that did not go out of their way to get consulting services (whereas in Ghana, we approached the businesses, not the other way around).
Thus a policy-relevant question remains: Would an expansion of this program have a similarly positive effect on those who did not self-identify as needing consulting advice? The effect could go up, or down, for those who do not seek the advice. In an earlier study on microenterprises in Peru, IPA found that the impact was higher on microentrepreneurs who expressed lower interest in business training. Thus the answer may not be as simple as providing the service to those who demand it the most.
We believe that economic development efforts are best served by testing and refining assumptions about what works, because despite the hopes and best intentions of smart people, not all interventions work. Finding different results in different contexts encourages us to look deeper into specific contexts and into the interventions themselves to determine which factors matter. Several of the differences in the chart on page 7 could explain the strikingly different results, but, in a sense, we have only two data points: one for Ghana and one for Mexico. In fighting poverty, simple questions typically have complex answers. This doesn’t mean the questions are unanswerable. It does mean that no one study, or in this case no two studies, will answer them all.